Accelerating microarchitectural simulation via statistical sampling principles
Abstract
The design and evaluation of computer systems rely heavily upon simulation. Simulation is also a major bottleneck in the iterative design process. Applications that may be executed natively on physical systems in a matter of minutes may take weeks or
months to simulate. As designs incorporate increasingly higher numbers of processor cores, it is expected the times required to simulate future systems will become an even greater issue. Simulation exhibits a tradeoff between speed and accuracy. By basing experimental procedures upon known statistical methods, the simulation of systems may be dramatically accelerated while retaining reliable methods to estimate error.
This thesis focuses on the acceleration of simulation through statistical processes. The first two techniques discussed in this thesis focus on accelerating single-threaded simulation via cluster sampling. Cluster sampling extracts multiple groups of contiguous
population elements to form a sample. This thesis introduces techniques to reduce sampling and non-sampling bias components, which must be reduced for sample measurements to be reliable. Non-sampling bias is reduced through the Reverse State Reconstruction algorithm, which removes ineffectual instructions from the skipped instruction stream between simulated clusters. Sampling bias is reduced via the Single Pass Sampling Regimen Design Process, which guides the user towards selected representative sampling regimens. Unfortunately, the extension of cluster sampling to include multi-threaded architectures is non-trivial and raises many interesting challenges. Overcoming these challenges will be discussed. This thesis also introduces thread skew, a useful metric that quantitatively measures the non-sampling bias associated with divergent thread progressions at the beginning of a sampling unit. Finally, the Barrier
Interval Simulation method is discussed as a technique to dramatically decrease the simulation times of certain classes of multi-threaded programs. It segments a program
into discrete intervals, separated by barriers, which are leveraged to avoid many of the challenges that prevent multi-threaded sampling.
Collections
Related items
Showing items related by title, author, creator and subject.
-
Ion Source And Sample Introduction Method And Apparatus Using Two Stage Ionization For Producing Sample Gas Ions
Eisele, Fred L.; Berresheim, Harald (6/8/1993)A high pressure interface device for introducing sample ions to a drift tube, an ion detection means or a mass mobility measurement means, having an ion source ionization region containing an isolated ionization source, a ... -
Challenges for the Heatshield Development of Sample Return Missions - An Overview on European Sample Return Studies and Requirements
Santovincenzo, A.; Ritter, Heiko; Walpot, L. (Georgia Institute of Technology, 2008-06-26)The atmospheric entry of the Earth return capsule of sample return mission is one of the most critical phases of sample return missions. The Earth return from extraterrestrial bodies (e.g. Mars, comets or asteroids) involves ... -
Summary report on sampling methods, sampling procedures, and a method for dirt analysis. Project 1850-8, final report to Jute Research Group
Institute of Paper Chemistry (Appleton, Wis.); Institute of Paper Chemistry (Appleton, Wis.) Jute Research Group. (Appleton, Wisconsin : the Institute,Georgia Institute of Technology, 1955-11-01)